Modelling Success of Commodity Futures: A Markovian Approach to Predict Volume

Abstract

Our study constructs an early and late stage for commodity futures and rigorously examine the determinants that contribute to success in each of these stages. It also studies the role of financialization as a determinant of commodity futures success. To assess our objective, we used a comprehensive dataset of 104 commodity futures from six exchanges in China, India, and the USA. The study reveals the factors influencing trading volume in commodity futures markets differ significantly depending on whether the contract is in its early or late stage. In the early stage, hedging effectiveness, competition, spot market volatility and Buyer concentration are crucial. However, in the Late Stage, financialization becomes a key driver as the market matures. Whereas the activeness of the cash market and homogeneity consistently boosts trading volume across both stages. Our study also employs Markovian models to predict the volume of commodity futures. This study is the first to employ Markovian models within a panel data framework, aiming to identify the best model for predicting volume. Finally, our findings demonstrate the superiority of the MFD-HMM model in predicting volume, as it proves to be the most effective for both in-sample and out-of-sample predictions.